<p>While highly educated talents are crucial for urban innovation and regional sustainable development, limited research has examined how environmental amenities affect their destination preference, particularly considering spatial heterogeneity and scale effects. Drawing on China’s 2020 Seventh National Population Census data across 337 cities, this study employs Multiscale Geographically Weighted Regression (MGWR) model to investigate the spatial patterns and determinants of highly educated talents in China, with a specific focus on environmental amenities. The results show that highly educated talents are primarily concentrated in provincial capitals, municipalities directly under the central government, separately administered cities, and resource-based cities in northern China. In contrast, southeastern China offers more favorable environmental amenities compared to the northwest, highlighting significant spatial disparities. Furthermore, the MGWR model outperforms the OLS and GWR models, with the Temperature-Humidity Index (THI) demonstrating a broader influence than PM_2.5 and the Normalized Difference Vegetation Index (NDVI). In addition, environmental amenities exhibit multilevel spatial heterogeneous effects on destination preferences of highly educated talents in China. Specifically, THI has a stronger positive effect in certain central and western cities. Our findings offer valuable insights for enhancing regional competitiveness through environmental quality improvements.</p>